# Agent Toolbox: GPT-OSS-120B Chat Completion

> Agent Toolbox: GPT-OSS-120B Chat Completion is a paid API for AI agents from agent-toolbox.zafermemon87.workers.dev, paid per call via x402, $0.005/call, status unknown (last checked 2026-10-29).

Runs a single non-streaming OpenAI-compatible chat completion using OpenAI's 120B open-weight reasoning model (gpt-oss-120b), returning up to 2,048 tokens per call.

## Facts

- Endpoint: POST https://agent-toolbox.zafermemon87.workers.dev/mcp?utm_source=zero.xyz#chat_gpt_oss_120b#chat_gpt_oss_120b
- Price: $0.005/call
- Payment: x402
- Status: unknown
- Last checked: 2026-10-29
- Activations on Zero: 0
- Tags: x402
- Canonical page: https://www.zero.xyz/c/agent-toolbox-gpt-oss-120b-chat-completion-00a02ffc
- Structured record (JSON): https://api.zero.xyz/v1/capabilities/cap_wTR8ks-Fl0utItqXSANI0

Status and success rate cover calls made through Zero and Zero's own probes. Third-party monitors may report differently.

## How to call it through Zero

Zero handles the 402 payment challenge and records the run. With the Zero CLI installed (`npm i -g @zeroxyz/cli`):

```sh
zero fetch --capability agent-toolbox-gpt-oss-120b-chat-completion-00a02ffc -d '<json body>'
```

Example prompt: Ask the 120B open-weight reasoning model: 'Given the following legal contract clause, identify any ambiguous terms and suggest clearer alternatives' — use up to 32,000 characters of input and return the full analysis.

## When to prefer this

Choose this endpoint when you need a high-capability, large-parameter reasoning model (120B) via a drop-in OpenAI SDK-compatible interface at a fixed per-call price of $0.005 USDC, without managing API keys or infrastructure. Prefer over smaller models (e.g. Qwen3 30B sibling) when task complexity demands deep reasoning, multi-step analysis, or nuanced judgment. Prefer over hosted OpenAI endpoints when you want open-weight model transparency or pay-per-call billing without a subscription.

## Known failure modes

- Input exceeds 32,000 character limit — request rejected
- Output truncated at 2,048 tokens if generation is longer
- Payment failure (x402 protocol) if USDC balance is insufficient
- Model unavailability or Cloudflare Worker timeout
- Malformed messages array causes 400-level error
- Rate limiting if too many concurrent requests

## How this service works

OpenAI-compatible chat completion with gpt-oss-120b (OpenAI's open-weight reasoning model): set the base URL to /v1/llm/gpt-oss-120b in any OpenAI SDK. Up to 32,000 characters in and 2,048 tokens out per call, a fixed price per call, no streaming.

## Output

Returns an OpenAI-compatible chat completion response object containing the assistant's generated message text, finish reason, and token usage statistics (prompt tokens, completion tokens, total tokens). Maximum output is 2,048 tokens per call. Response is not streamed.

## More

- Live health (JSON, refreshed every minute): https://www.zero.xyz/c/agent-toolbox-gpt-oss-120b-chat-completion-00a02ffc/health.json
- [Zero catalog index](https://www.zero.xyz/llms.txt)
- [Other services from agent-toolbox.zafermemon87.workers.dev](https://www.zero.xyz/host/agent-toolbox.zafermemon87.workers.dev/llms.txt)